Skaters have no impact on their goaltenders performance

[abo9]

The pressured forward generates less xGA.

The rebound shots are classified, extra xGA is added for them.

All those cases impact xGA, not the f(GA,xGA) that Micah is measuring as goaltender's performance. He classifies those xGA changes as 'defensive impact of a player'.

Except the last sentence. According to Micah and his findings similar shots don't, after all, differentiate. His findings indicate function f(GA, xGA) of a given goalie doesn't change because of the corps in front.

As I said, there's a perception catch. If the D allowed for 15 high danger chances which resulted in 9 goals vs the defense that allowed 5 high danger chances where the result was 3 goals, the G performance might as well be the same. Micah didn't find a consistent case where similar 5 high danger chances same goalie faced would result in 3 goals with players X,Y,Z on the ice and in 1 goal with players A,B,C.

Gotcha thanks. How do we know that a forward is pressured?

Or do you mean that the difference would be captured in XGa if it had a measurable impact?

I personally find it hard to believe that we can measure a goalie's performance irrespective of players in front of them. It seems like the goalies playing for good teams consistently float to the top - unless I'm misunderstanding the whole point that is being made.

I guess theres also the case where goalies impact team performances by a lot?

 

[lomiller1]

If you think the first 5 items "do not correlate significantly" with the defensive corps on the ice, ... well, there's a reason why I call much of what people are doing these days teh analytics. And no, I don't mean that in the nicest way possible, either.

2 (shot location and tips\deflections) of those 5 are tracked and factored into xGoals. There is little evidence velocity matters in any significant way, but if it does, I don't see how the defender controls it. There is a suspicion some passes matter, but but as of yet this can only be captured in subjective ways. Again though, can a defender really control how crisp an offensive players passes are? It's been demonstrated that one-timers matter but they are not tracked NHL which records them as slap shots, so it's a known gap but again not nessiarily a gap wrt defensive play.

Even when there are gaps in xGoals it doesn't follow that defenders can influence them. Indeed crisp passing, slap shots and shot velocity are all clearly offensive skills. What McCurdy is showing is mathematical evidence that on aggregate the "gaps" in xGoals are not statistically significant factors in defensive performance and that treating them as purely random doesn't reduce xGoals ability to capture defensive skill.

This in turn would suggest there in little or no opportunity to improve your teams defensive play by focusing on things not already included in xGoals. It is however conceivable that every team already does these things equally well, in which case slacking off to much could still hurt your defensive performance. More likely, however, is that these are just offensive skills defenders can't take away effectively.

 

[lomiller1]

Gotcha thanks. How do we know that a forward is pressured?

Or do you mean that the difference would be captured in XGa if it had a measurable impact?

I personally find it hard to believe that we can measure a goalie's performance irrespective of players in front of them. It seems like the goalies playing for good teams consistently float to the top - unless I'm misunderstanding the whole point that is being made.

He's not saying defensive players don't impact a goaltenders GAA. What he's saying is that the ways in which they impact a goaltenders GAA is already well characterized by xGoals and that the things xGoals can't yet capture wrt defensive players don't impact a goaltenders GAA in a way that can be distinguished from random chance.

 

[lomiller1]

The idea that GA shouldn't be used to measure a skater's defensive ability? Hogwash. It's not a great stat, it's a poor measure, but like all statistics it requires proper interpretation to understand when it's useful and when it's not. That's an exercise for individuals, and that can be done properly without someone generating 667GB of data to create models with 374 possible variables that need to run for 18.26 hours to come up with "brilliant" conclusions.

GA by itself is a counter indicator in many, perhaps most cases, eg Defender A plays 5 min a night and defender B plays 25 min a night. Which one is likely to be better defensively? Which one is likely going to give up more goals?

More importantly which one is likely to be more effective defensively? A bad defensive player who spends spends significantly more time in the offensive zone because they retrieve lose pucks and dominate zone exits can still give up less than a good defensive player who struggles with zone exits. But, they may not. If you think you can keep track of enough information in your head to make this determination without the use of statistics and analytics you are fooling yourself.

 

[abo9]

He's not saying defensive players don't impact a goaltenders GAA. What he's saying is that the ways in which they impact a goaltenders GAA is already well characterized by xGoals and that the things xGoals can't yet capture wrt defensive players don't impact a goaltenders GAA in a way that can be distinguished from random chance.

I reread his tweets, just trying to understand better, but what sticks out to me is... idk how he characterizes "defensive prowess". Is it xga, which would be redundant, or stuff like actual positioning (I somehow doubt it, but ifk his work enough).

ahhhh, I think I got a flash of understanding... I'm probably just rephrasing what others have said, but for my own benefit lol.

It's not that skater performances don't impact goalie performances, its that shot quality or whatever its called in his model, already takes the skater's performance into account - a defenseman being out of position or having poor stick work would allow for better/more dangerous shots to be taken.

But yeah, I guess that means it doesnt impct the goalie's performance at an expected saves above expected level.

Smaller details that are unaccounted for would be lumped into the error term (randomness) - the small % that analysts will work extra hard to get insights into to get a competitive advantage.


What still trips me is that, when taking a (admittedly) quick look at Moneypuck's data on goals saves'd above expected, the ranking of teams year to year seems more consistent than the ranking of goalies. Ie.Boston Islanders Arizona Carolina Nashville have seen quite a few goalie changes over the years, and their goalies frequently appear on top.

Wouldnt that suggest that something about the skaters help goalies have a better xgaa? That maybe, whats unaccounted for isnt "small details" to be glossed over? Just raising the question, year to year rankings seems better correlated with teams than goaltenders. Micah definitely put more efforts than me onto it, but again, trying to understand better whats being discussed here.

Another reasoning would be that we unpacked goalie performance really well, but that theyre all just too close in talent to one another and you'd be just better off to pick the cheapest option?

 

[morehockeystats]

Gotcha thanks. How do we know that a forward is pressured?

Or do you mean that the difference would be captured in XGa if it had a measurable impact?

I personally find it hard to believe that we can measure a goalie's performance irrespective of players in front of them. It seems like the goalies playing for good teams consistently float to the top - unless I'm misunderstanding the whole point that is being made.

I guess theres also the case where goalies impact team performances by a lot?

Well, Micah's research shows that, simply put, there are no NHL players that consistently:

• trip their goalies or push forwards into them
• park themselves in their line of sight
• allow more than usual opposition into their line of sight
• make shooters whiff on their shots

And as I said earlier, the research is easily falsifiable, i.e. there is a possible dataset that refutes it, the question is whether it actually exists. But once produced, it will put this argument to reset.

 

[ForsbergForever]

So does the essentially mean that if I had Ray Bourque and Nicklas Lidstrom on defense, the result would be exactly the same as having Tyler Myers and Rasmus Ristolainen on the ice instead?

 

[Bear of Bad News]

So does the essentially mean that if I had Ray Bourque and Nicklas Lidstrom on defense, the result would be exactly the same as having Tyler Myers and Rasmus Ristolainen on the ice instead?

It means that the headline is misleading (more to the point, it means that you need to carefully define what you mean by "goaltender's performance").

 

[dr robbie]

Zero face validity. Play hockey for 5 minutes and you'll know otherwise.

If your data is demonstrating something else, your data is flawed or your tracking the wrong data.

 

[Bear of Bad News]

Zero face validity. Play hockey for 5 minutes and you'll know otherwise.

If your data is demonstrating something else, your data is flawed or your tracking the wrong data.

Or you're using a different definition of "goaltender's performance" than the study is using.

Many of the people discussing this (myself included) have played hockey for five minutes. Your appeal to authority isn't helpful.

 

[Panthaz89]

Bad defense can't wear a goalie out?

 

[JaegerDice]

I dont understand how Micah’s conclusion is in any way controversial… of course the goalie alone makes the difference between xGA and GA…

So does the essentially mean that if I had Ray Bourque and Nicklas Lidstrom on defense, the result would be exactly the same as having Tyler Myers and Rasmus Ristolainen on the ice instead?

No, because Lidstrom and Borque would significantly impact the number of expected goals against by limiting shots and chances.

Dmen can limit opportunities, but they cant make the saves for the goalies.

 

[Bear of Bad News]

I dont understand how Micah’s conclusion is in any way controversial… of course the goalie alone makes the difference between xGA and GA…

If I'm understanding the study, he's basically saying that his model does a good job of isolating that component.

 

[Bear of Bad News]

Bad defense can't wear a goalie out?

I've also heard the opposite - that great defense makes a goaltender complacent.

I don't think either claim has been exhibited in data.

 

[ForsbergForever]

I've also heard the opposite - that great defense makes a goaltender complacent.

I don't think either claim has been exhibited in data.

Playing in front of great defense didn't seem to hurt Dryden or Roy, for example. Is there any evidence for the complacency idea?

 

[Bear of Bad News]

Playing in front of great defense didn't seem to hurt Dryden or Roy, for example. Is there any evidence for the complacency idea?

I was merely offering it as the counter to "bad defenses make goaltenders worse".

The standard argument involves lower save percentages, but of course, that's usually mitigated by things that xGA accounts for.

 

[Michael Farkas]

I also don't think there's a singular answer there because goalies are wired differently (even intra-species)...like, I wonder aloud about Patrick Roy playing behind some good defenses and, frankly, I don't think he gets enough discredit for some really putrid goals that he's given up in the playoffs over the years (of course, bound to happen to anyone when you play two zillion playoff games, but...) - I wonder if there's any tie to, say, not seeing a scoring chance for six minutes and then Cam Neely blows one by him from 45 feet...

Having seen the vast majority of Fleury's career, he certainly seemed to be more engaged when he was under siege against all odds...like game 5 of the 2008 Final, for instance...but he's notoriously not focused when he doesn't see pucks. In fact, I'd bet a round of drinks on Fleury being the all-time leader in giving up goals on the first shot of the game...at least in my lifetime...I don't know if Murray Bannerman or someone from the lean years is maybe guilty by association (era) here, but Fleury couldn't bothered to be hit with one early in games...

Regardless, over enough time, I don't think any goalie's performance would drop because he had a great defense...ancillary stuff like save pct might be affected, but as is well established on this board, I think save pct. can go kick rocks...

 

[lomiller1]

I was merely offering it as the counter to "bad defenses make goaltenders worse".

The standard argument involves lower save percentages, but of course, that's usually mitigated by things that xGA accounts for.

Martin Broudeur comes to mind as someone who frequently had his mediocre save % explained away as being the result of him facing very few shots.

 

[Michael Farkas]

But I mean, that's the reality, right? It's not like he was giving up more goals and that's why his save pct is low...everyone gives up 2. If you give up 3, no one is talking about you. If you give up 1, there's no debate. So who cares if you give up 2 on 17 or 2 on 39...?

 

[morehockeystats]

But I mean, that's the reality, right? It's not like he was giving up more goals and that's why his save pct is low...everyone gives up 2. If you give up 3, no one is talking about you. If you give up 1, there's no debate. So who cares if you give up 2 on 17 or 2 on 39...?

17 shots can produce more xGA than 39.

 

[morehockeystats]

As I said before, the study demonstrated that the defencemen bad enough to consistently obstruct their own goalies, or to consistently allow opposition forwards to obstruct their goalies, do not make it to the NHL.

I've also heard the opposite - that great defense makes a goaltender complacent.

I don't think either claim has been exhibited in data.

probably evens out

 

[morehockeystats]

Gotcha thanks. How do we know that a forward is pressured?

Or do you mean that the difference would be captured in XGa if it had a measurable impact?

I personally find it hard to believe that we can measure a goalie's performance irrespective of players in front of them. It seems like the goalies playing for good teams consistently float to the top - unless I'm misunderstanding the whole point that is being made.

I guess theres also the case where goalies impact team performances by a lot?

A pressured forward would usually produce a different shot type, e.g. a poke instead of a wrister. The forward would also usually drift to a less favorable location.

John Gibson had been described to have stellar performances even when Anaheim started to be real bad.

 

[morehockeystats]

Martin Broudeur comes to mind as someone who frequently had his mediocre save % explained away as being the result of him facing very few shots.

NHL goalies better vs high shot volumes

I'll just pick this out for now as the response is fairly straight forward. I may respond to the rest later. I'd love to hear your response. I'm not talking about games where a goalie *did* face a large number of shots and got pulled, I'm talking about games where a goalie *would have* faced...

forums.hfboards.com

 

[bucks_oil]

I also don't think there's a singular answer there because goalies are wired differently (even intra-species)...like, I wonder aloud about Patrick Roy playing behind some good defenses and, frankly, I don't think he gets enough discredit for some really putrid goals that he's given up in the playoffs over the years (of course, bound to happen to anyone when you play two zillion playoff games, but...) - I wonder if there's any tie to, say, not seeing a scoring chance for six minutes and then Cam Neely blows one by him from 45 feet...

Having seen the vast majority of Fleury's career, he certainly seemed to be more engaged when he was under siege against all odds...like game 5 of the 2008 Final, for instance...but he's notoriously not focused when he doesn't see pucks. In fact, I'd bet a round of drinks on Fleury being the all-time leader in giving up goals on the first shot of the game...at least in my lifetime...I don't know if Murray Bannerman or someone from the lean years is maybe guilty by association (era) here, but Fleury couldn't bothered to be hit with one early in games...

Regardless, over enough time, I don't think any goalie's performance would drop because he had a great defense...ancillary stuff like save pct might be affected, but as is well established on this board, I think save pct. can go kick rocks...

As a member of "a different species" you can take this personal experience for what it's worth. I'm a "B-level" (according to standard US adult hockey ratings) beer-leaguer goalie, that converted to hockey after being a higher level soccer goalie. I played only a bit of hockey growing up in Canada (once punched Pronger!), but quit for soccer at a young age.

Most of my experience is as an adult in a draft league, where every 4 months we re-draft the teams. Players are mostly high-school varsity with a very few ex-college/ex-junior A level players. The redrafting of teams let's me play on three teams per year and experiment with different team construction (I'm often the drafting "captain" of my team).

Here is what I've found and chatted with many goalie friends, so there is some truth in here:

1) Getting the puck out (or in at the other blue) matters to your goalie. If you have a chance to get the puck out and don't... I'll stop it for you. But if you keep making the same mistake over and over... it's gonna piss me off, and I can't concentrate when I'm pissed off. I can't help it... and I'm not a pro... but I pride myself on "staying in the game" and all goalies pride themselves on "no mistakes/attention to detail", but if it is obvious to me that you aren't paying attention to detail, a sinister part of me will say "well you deserved it" as a somewhat stoppable shot squeaks in.

2) Trust is built in two directions. I've had solid, stay at home and two-way defenders that were prone to a huge blunder every once in a while. These are the same guys who clear pucks off my line or press them back into/under my pads for a whistle. When these guys blow a tire, I'm more likely to make a save... it's just the way it is. I know they've got my back and my adrenaline will kick in... gotta make a save for Jonesy... must make this save! That adrenaline is like confidence and I'll stay out higher, cut down the angle, never shrink back into the net... and the save will be made if it is anywhere within my physical capability. The same situation with a selfish winger who never passes the puck and turns it over at the blue line? Same thing, I'll stop it for my teammates, and for myself, but he better give me a stick tap. The third time he does it that night? I want to do the same, I want to win, but my brain goes "WTF, not again" and then I don't get that adrenaline rush, I feel frustration, and my SPCT in those situations is lower... no stats... I just know it is.

3) Confidence matters, saves build confidence: To your point... Fleury under siege has just made 3-4 five-bell saves. He KNOWS it's not him. He KNOWS he's helping his teammates. That breeds confidence, action gives adrenaline, games like that are a constant flow of positive energy... until physical exhaustion or frustration with teammates set in. The latter almost NEVER happens in a 2-1 type of game, but if I'm under siege in a 5-4 game where I'm seeing way more action than the other guy, then my hot streak isn't going to last as long... frustration will set in... and then see point #1 and #2 above.

4) Sometimes, playing for strong teams can lower SPCT: but probably not for the reasons you are thinking. On a strong possession team, you will get less shots... and you'll be ready, that's your job. BUT... what also tends to happen on strong possession team is that the shots you DO get are higher quality. Everybody is human and a team of strong defenders will occasionally still give up a breakaway when a pass at the far blue line is intercepted. At the NHL level a breakaway (see shootout stats) is ~50% SPCT.... so obviously that's going to lower your SPCT on that night. Same with a 2 on 1. Same with a team that blocks a tonne of shots, the one that gets through is probably screened, or tipped, or... you see my point. SPCT can go down in these situations, even though the goalie is stopping all of the stoppable shots.

5) But often it works the other direction: A truly strong defending team, one that isn't prone to huge blunders, will improve your SPCT overall. Shooters are pressured on their shots, they won't have time to pick their spot, sticks are tied up, crease area is cleared so I can see, shots then come from the perimeter, etc... all of those fundamental plays shoot my SPCT through the roof. It becomes a super easy night. Ultimately it's attention to details and details are what prevents goals. Goalies live it, defensive minded D-men and centers know it, plenty of forwards and offensive D can be taught.... but one-dimensional wingers... just can't be helped. Never pick (more than one of... and only if you must) those guys

 

[Bear of Bad News]

Agreed with all of the above and better said that I certainly could do, noting that this thread is specifically about impact on xGA (or lack thereof) and not on save percentage, GAA, or wins.

80% of this thread's problems stem from a wholly inaccurate subject line.